[ntp:questions] OT: GPS18x LVC failure

David Lord snews at lordynet.org
Mon Oct 12 10:04:08 UTC 2009


David J Taylor wrote:
> "David Lord" <> wrote in message news:7jen4sF32s09iU1 at mid.individual.net...
> []
>> The pps via usb doesn't work with NetBSD. That is I believe lack
>> of hardware interrupt capability via usb. Possibly latest
>> development pps might work from usb on Linux.
> 
> It certainly works on Windows.  Is that because Windows does a better 
> job of emulating the interrupts than NetBSD does, or because of the 
> architecture which Dave Hart developed for Windows does a better job of 
> picking up the timing, or because I was lucky with my USB device or its 
> drivers?

Ntpd on Linux allows an extra mode to what's usable on version
installed with NetBSD even though I see that particular syscall
now appears to be available on NetBSD-5. I think it's just ntpd
version lagging behind but I don't have time to play with
development branch. That mode with radioclkd2 is claimed not to
be too far from pps for MSF/DCF77.

.....

> 
> If the Garmin can deliver +/-5V, it should certainly drive almost any 
> reasonable RS-232 port.  It sounds as if your USB converter has a lower 

I doubt now that I tried with the Garmin on the server (possibly
tried on eeepc but not yet sorted the data). I'd had <165us at 50%
from MSF/SHM/PPS via serial, then just after that noted failure
to get useful result via usb-serial, no better than 10ms. I don't
have figures handy for MSF without PPS via serial. Outputs for
MSF/DCF77 are current sinks to 0V, Conrad pulldown for MSF and a
BC337 on my PPS module.

David




More information about the questions mailing list